Instrument Presentation

Laser wavelengths ranging from ultra-violet through visible to near infra-red can be used for Raman spectroscopy. Typical examples include (but are not limited to):

• Ultra-violet: 244 nm, 266 nm, 320 nm, 355 nm, 405 nm
• Visible: 458 nm, 473 nm, 515 nm, 532 nm, 594 nm, 633 nm, 660 nm
• Near infra-red: 785 nm, 830 nm, 980 nm, 1064 nm

The choice of laser wavelength has an important impact on experimental capabilities:

Sensitivity

Raman scattering intensity is  proportional  to  λ^-4  where λ  is the laser wavelength. Thus an infra-red laser results in a decrease in scattering intensity by a factor of 15 or more, when compared with blue/green visible lasers.

Spatial resolution                      

The diffraction limited laser spot diameter can be calculated according to the equation, diameter = 1.22 λ/NA (where λ is the wavelength of the laser, and NA is the numerical aperture of the microscope objective being used). For example, with a 532 nm laser, and a 0.90/100x objective, the theoretical spot diameter will be 0.72 µm – with the same objective, a 785 nm laser would yield a theoretical spot diameter of 1.1µm. Thus, achievable spatial resolution is partially dependent on choice of laser.                                                                                                                                                                                                           

Optimization of resulting based on sample behavior

For example:

• Blue or green lasers can be good for inorganic materials and resonance Raman experiments (e.g., for carbon nanotubes and other carbon materials) and Surface Enhanced Raman Scattering (SERS).
• Red or near infra-red (660-830 nm) are good for fluorescence suppression.
• Ultra-violet lasers for resonance Raman on bio-molecules (such as proteins, DNA, and RNA), and fluorescence suppression.

Ultra-violet (UV) lasers for Raman spectroscopy typically include laser wavelengths ranging from 244 nm through to 355 nm.

Theoretically UV Raman spectroscopy is not different from standard analysis using visible laser wavelengths. However, in practice there are a number of practical difficulties and disadvantages which must be considered.

Advantages

• With certain samples, UV laser excitation   can interact in ways not possible when using visible laser sources. For example, in semiconductor materials, the penetration depth of UV light is typically in the order of   a few nanometers, and thus UV Raman can be used to selectively analyze from a thin top surface layer (as is commonly found in silicon on insulator SOI materials). In another example, UV excitation can give rise to specific resonance enhancement with biological moieties, particularly protein, DNA and RNA structures. Specific analysis of these materials within tissue can be difficult using visible laser wavelengths.

• Fluorescence suppression can often be assisted using UV lasers, by spectrally separating the Raman and fluorescence signatures. With visible lasers it is common that Raman and fluorescence are superimposed, and the incomparable strength of the fluorescence is what can perturb (or completely mask) the Raman spectrum. With UV excitation, the Raman spectrum lies close to the laser line, whereas the fluorescence is often slightly removed to higher wavelengths. Thus, they no longer overlap, and the fluorescence is no longer an issue.

• Increased sensitivity can result from UV excitation, since Raman scattering efficiency is proportional to λ^-4 , where λ  is the laser wavelength. Thus, Raman scattering at 325 nm is a factor of 14 times more efficient than that at 633 nm.

Disadvantages

• UV Raman still remains a more sophisticated technique which requires greater expertise to handle. Reasons for this include the fact that the laser beam is now invisible, and that the lasers are larger, more complex, and considerably more expensive.

• Samples are more prone to burning and degradation from the laser beam since the energy per photon is increased. However, new techniques such as DuoScan™ optics allow the laser beam to be rapidly rastered over the sample and thus prevent immediate burning. As an example, cellulose will burn with 325 nm excitation within a few milliseconds, but with DuoScan™ it remains resilient to burning for more than five minutes.

• Many Raman systems designed for visible and near infra- red analysis are not suitable for UV Raman. UV Raman requires specific mirror coatings, microscope objectives, diffraction gratings, and a CCD detector for optimized results. Modern systems such as the LabRAM HR can be configured to work efficiently from the UV through to the infra-red without compromise, but nonetheless the additional requirements do come at a cost.

Near infra-red (NIR) lasers for Raman typically include a range of wavelengths greater than 700 nm, such as 785 nm, 830 nm, 980 nm and 1064 nm. The key reason for the use of NIR Raman is for fluorescence suppression, but there are    a number of drawbacks which must be considered. Whilst at times NIR Raman is invaluable, it should certainly not be considered the best solution for every sample.

Advantages

• Fluorescence suppression can often be assisted using NIR lasers. Fluorescence is a two photon process, which first requires absorption of a photon, and is followed by the emission of a fluorescent photon. On the other hand, Raman is a one photon scattering process, which does not require absorption. While many materials absorb in the visible region, fewer do so in the NIR region.Thus, in many cases NIR lasers will not give rise to fluorescence (since absorption does not occur), but Raman scattering will be present as normal. In cases where samples strongly fluoresce with visible excitation NIR Raman can provide a solution and allow a good Raman spectrum to still be obtained.

Disadvantages

• Decreased sensitivity can result from NIR excitation, since Raman scattering efficiency is proportional to λ^-4 , where λ is the laser wavelength. Thus, Raman scattering at 785 nm is almost a factor of 5 times less efficient than that at 532 nm. This problem is compounded by decreasing sensitivity of the CCD detector in the NIR Raman range. As a result, measurement times are often considerably increased to get similar spectral quality to measurements made with visible lasers.

• NIR lasers often have beam qualities (e.g., beam width and divergence) which are not as well suited for microscopy. The result is that spatial resolution can be somewhat compromised, and thus achievable results may not meet theoretical predictions.

There are two main classes of filters used for Raman spectroscopy.

Optical filters

These optical components are placed in the Raman beam path, and are used to selectively block the laser line (Rayleigh scatter) while allowing the Raman scattered light through   to the spectrometer and detector. Each laser wavelength requires an individual filter. There are two main types of filters used, both of which can be used without user intervention or optimization:

• Edge. An edge filter is a long pass optical filter which absorbs all wavelengths up to a certain point, and then transmits with high efficiency all wavelengths above this point. As an example, a 532 nm edge filter will absorb all light up to 533.5 nm or above (e.g., 50 cm^-1 or above) including the laser emission. Above 533.5 nm it will transmit light, allowing detection of the Raman spectrum (stretching from 50 cm^-1 up to 3500 cm^-1 or  above).  The edge filters have an ultra steep edge between the absorbing and transmitting spectral region, and offer excellent blocking of the laser line.The advantage of the edge filter is that it is environmentally stable with a near infinite lifetime.

• Holographic notch. A notch filter has a sharp, discrete absorption which for Raman is chosen to coincide with a specific laser wavelength. Typically the absorption will be a few nanometers wide (corresponding to a few hundred cm^-1). As an example, for a 532 nm laser, a notch filter will be chosen with a central absorption at 532 nm. The laser line will be absorbed, but the Raman spectrum above will be transmitted. Unlike the edge filter, the notch does have a finite lifetime, and will degrade with time. The advantage of the notch is that it allows measurements to be made for both the Stokes and anti-Stokes Raman scattering, which is useful for certain specialized measurements.

• Tunable filtering spectrometer with a triple monochromator instrument, it is possible to work in a ‘double subtractive, single spectrograph’ mode. The first two monochromators are used as an integral pair, which first disperses both the Rayleigh (laser) and Raman scattered light, physically blocks the Rayleigh, and then recombines the light. The third spectrometer then acts in the normal way to disperse the Raman scattered light with subsequent detection. The advantage of using a triple spectrometer in this way is that the laser filter is infinitely variable, and the one instrument can be used to work with any number of laser sources. In addition, the filtering performance is excellent, and allows Raman analysis down to 4-5 cm^-1. However, such an instrument does require rather more expertise to operate compared to the more standard filter based single monochromator systems, and therefore would rarely be considered for routine analysis.

A CCD (Charge Coupled Device) is a silicon-based multichannel array detector of UV, visible and near-infra light. They are used for Raman spectroscopy because they are extremely sensitive to light (and thus suitable for analysis of the inherently weak Raman signal), and allow multichannel operation (which means that the entire Raman spectrum can be detected in a single acquisition). CCDs are widely used, not least as the sensors in digital cameras, but versions for scientific spectroscopy are of a considerably higher grade to give the best possible sensitivity, uniformity and noise characteristics.

CCD detectors are typically one dimensional (linear) or two dimensional (area) arrays of thousands or millions of individual detector elements (also known as pixels). Each element interacts with light to build up a charge – the brighter the light, and/or the longer  the  interaction,  the  more  charge is registered. At the end of the measurement read out, electronics pull the charge from the elements, at which point each individual charge reading is measured.

In a typical Raman spectrometer, the Raman scattered light is dispersed using the diffraction grating, and this dispersed light is then projected onto the long axis of the CCD array. The first element will detect light from the low cm^-1 edge of the spectrum, the second element will detect light from the next spectral position, and so on...the last element will detect light from the high cm^-1 edge of the spectrum.

CCDs require some degree of cooling to make them suitable for high grade spectroscopy. Typically this is done using either peltier cooling (suitable for temperatures down to -90ºC), and liquid nitrogen cryogenic cooling. Most Raman systems use peltier cooled detectors, but for certain specialized applications, liquid nitrogen cooled detectors still have advantages.

An Electron Multiplying CCD (EMCCD) is a special type of CCD detector, which uses the latest technology to enhance the spectrum quality when extremely low signal levels are present. This enhancement is particularly valuable when the Raman signal is very weak, since the electron multiplication process can result in good spectrum quality, unlike the conventional CCD where only a few of the stronger features can just be observed above the noise. The benefits of EM gain are clearly obvious in fast Raman spectral imaging,where the necessary short integration times can often result in signals which are barely visible above the noise when measured with a conventional CCD.

The EMCCD has two readout registers on the chip – a conventional register and an electron multiplying (EM) register. In the EM register, the clocking voltages used are higher than for conventional clocking, causing the electrons to acquire sufficient energy that impact ionization can occur. At this point, extra electrons are produced and stored in   the next pixel. There is only a small probability of electrons acquiring sufficient energy for impact ionization to occur (thus creating additional electrons), but since the readout register has many elements within it, significant gain factors are possible (up to ~1000x). The key benefit of an EMCCD is that the amplification occurs before readout of the signal, which means that the signal is not readout noise limited. In other words, through amplification, the signal is raised well above the noise floor which is largely determined by the noise of the readout electronics (pre-amplifier and A/D converter).

Spectral resolution is the ability to resolve spectral features and bands into their separate components.The spectral resolution required by the analyst or researcher depends upon the application involved. For example, routine analysis for basic sample identification typically requires low/medium resolution. In contrast, characterization of polymorphs and crystallinity often requires high resolution, since these phenomena exhibit only very subtle changes in the Raman spectrum, which would not be visible in a low resolution experiment.

Spectral resolution is an important experimental parameter. If the resolution is too low, spectral information will be lost, preventing correct identification and characterization of the sample. If the resolution is too high, total measurement time can be longer than necessary. What makes resolution “too low” or “too high” depends upon the particular application, and what information is desired from the experiment.

Typically, low/medium resolution is suitable for basic chemical identification, and distinguishing different materials. Higher resolution becomes necessary to characterize more subtle spectral features – for example, minor changes in the shape or position of a peak. There are a number of chemical phenomena which cause such subtle spectral changes:

• Crystallinity – In general Raman peaks become sharper and more intense as a material moves from an amorphous to crystalline structure. High spectral resolution allows even very small changes in crystallinity to be characterized.

• Polymorphism – Polymorphs are materials which have the same chemical formula but differing solid state forms. Because the chemical formula is identical, their overall Raman spectral profiles are similar, but the influence of the solid form on individual bond vibrations causes subtle changes in the spectrum. Thus, it is not unexpected to observe minor changes in peak shapes and positions throughout the spectrum.

• Intrinsic stress/strain – Some materials (most notably semiconductors) display changes in their Raman/ photoluminescence spectra when subjected to stress and strain. In the case of  semiconductors,  stress/ strain is carefully induced within the material to produce semiconducting and/or luminescent properties necessary for use of that material in working devices.Raman is used as a vital tool to characterize the stress/strain, but often the effect on the spectrum is subtle – thus, high resolution is necessary.

• Hydrogen bonding – The weak inter- and intramolecular interactions caused by hydrogen bonding can cause small changes in a Raman spectrum, and with high resolution capability, the spectrum can be used to investigate such interactions.

• Protein folding – A protein’s primary structure will, of course, have a significant impact on the resulting Raman spectra, but the secondary and tertiary structures comprising localized and more widespread folding cause sufficient perturbation to the vibrational modes to affect the spectrum. Once again, the effect will be subtle, and only high resolution analysis will allow its characterization.

Spectral resolution in a dispersive Raman spectrometer is determined by four main factors. In the discussions below, the effect of each factor is considered under the assumption that all other factors remain unchanged. In real life all of these factors can exist in many varied permutations, which makes direct comparison of a system’s performance and capabilities difficult.

Spectrometer focal length

The longer the focal length (e.g., the distance between the dispersing grating and detector) of the spectrometer, the higher the spectral resolution. Typical Raman spectrometers have focal lengths ranging from 200 mm (for low/medium resolution) through to 800 mm and higher (for high resolution). It is sometimes forgotten that a long focal length spectrometer is not limited to high resolution work only – With a suitable choice of gratings (see below), a high resolution spectrometer can be run in a low reso-lution mode. In this way, it is ideally suited for low/medium resolution analysis for routine screening, and yet can also offer high resolution analysis for more specialized applications.

Diffraction grating

The higher the groove density of the grating (typically measured as number of grooves per millimeter), the higher the spectral resolution. Typical gratings used for Raman vary from perhaps 300 gr/mm (low resolution) through to 1800 gr/mm (high resolution). More specialized gratings (including 2400 gr/mm and 3600 gr/mm) are also available, but have certain limitations, and should not be considered general purpose. The use of higher groove density gratings cannot be applied ad infinitum to increase spectral resolution, since they will have fixed practical and physical limits linked with the spectrometer itself. Thus, gratings provide an initial way to improve resolution, but once their limit is reached, it is necessary to move to a longer focal spectrometer.

Laser wavelength

The dispersing power of a grating/spectrometer pair can usually be considered constant in terms of wavelength. However, Raman spectra use an energy related unit (Raman shift, or wavenumber, cm^-1) which means that the spectral resolution decreases as the laser excitation is changed from infrared to visible to ultra-violet wavelengths. As an example, if a 600 gr/mm grating is used with an infrared laser, a 1200 gr/mm or 1800 gr/mm will be required with a green laser to achieve a similar resolution.

Detector

Most systems have a single detector, so practically the user does not have control of this factor. However, it should be noted that different detectors can be configured with different pixel sizes. The smaller the pixel the higher the achievable spectral resolution.

A Raman microscope combines a Raman spectrometer with a standard optical microscope. The excitation laser beam   is focused through the microscope to create a micro-spot with a diameter in the order of 0.5-10 µm. The Raman signal from the sample is collected from a similar area, passes back through the microscope into the spectrometer and is there analyzed for spectral information.

The Raman microscope allows Raman spectroscopy to be performed with microscopic spatial resolution. Thus it opens up a new dimension in chemical analysis:

• Analysis and identification of individual particles with dimensions down to 0.5 µm.
• Characterization of sample features with dimensions down to 0.5 µm.
• Location and identification of microscopic contaminants.
• Raman mapping (imaging) of sample features to show distribution of components, with a spatial resolution down to 0.5 µm.

Simply adding a microscope assists in giving lateral (XY) spatial resolution, but does not give depth (Z) spatial resolution. For this confocal optics are required. There are several methods in use today, some truly confocal, others pseudo confocal, which work with varying success. For a true confocal design (which incorporates a fully adjustable confocal pinhole aperture) micron depth resolution is possible, allowing individual layers of a sample to be discretely analyzed.

Some Raman microscopes do not have confocal optics. Simply adding a microscope assists in giving lateral (XY) spatial resolution, but does not give depth (Z) spatial resolution. For this, confocal optics are required. There are several methods in use today, some truly confocal, others pseudo confocal, which work with varying success. With a true confocal design (which incorporates a fully adjustable confocal pinhole aperture) micron depth resolution is possible, allowing individual layers of a sample to be discretely analyzed.